Persons with schizophrenic disorders have significantly higher mortality rates than persons in the general population. It is likely that modifiable patient behaviors and health care delivery system characteristics contribute to this phenomenon. Identification of such modifiable patient and health care system factors may be critical to improving the health status and longevity of persons with schizophrenia. Due to the complexity of this issue, we selected a "tracer" condition strategy, focusing on a single medical disease, diabetes, as a prototypical serious, chronic medical problem from which lessons may generalize to other disorders. Diabetes is a highly prevalent chronic medical disorder which requires active self-care and which may serve as a prototype disorder to understand problems that persons with schizophrenia and other severe mental illnesses have in obtaining and making use of adequate medical care. Using the health beliefs model as a conceptual framework, this study will compare 100 persons with schizophrenia who have diabetes to 100 persons with major depression and to 100 persons without serious and persistent mental illness who suffer from diabetes on: 1) diabetes-specific health behaviors (e.g., compliance), health outcomes (control of blood sugar), and quality of diabetes care received; 2) potential mediating variables, hypothesized to differ between schizophrenics and non-schizophrenics, including diabetes health beliefs and illness knowledge. Additional analyses will compare the schizophrenia sample to published norms on diabetes-specific measures. We will also examine whether system characteristics (e.g., integration of physical care and psychiatric care) and clinical characteristics (e.g., social skills and symptoms) account for poorer outcomes among persons with schizophrenia and major depression. Patients with schizophrenia and major depression will be drawn from a variety of mental health centers in the Baltimore area. A matched sample will be drawn from a University of Maryland Family Practice Clinic. Patients will receive a single assessment and chart review. This proposal thus focuses on diabetes and schizophrenia, although our underlying assumption is that our findings will contribute to our understanding of management and outcomes of other medical illnesses in persons with schizophrenia.